Carcinoembryonic antigens: alternative splicing accounts for the multiple mRNAs that code for novel members of the carcinoembryonic antigen family

A monoclonal anti-idiotypic antibody bearing the image of an epitope specific to the human carcinoembryonic antigen

One concept for immune therapy of patients bearing carcinomas involves monoclonal anti-idiotypic antibodies (Malds) to trigger the immune system of the host into a response against the tumor cells. Current theory states that so-called internal image Malds bearing epitopes specific to a given tumor-associated antigen would be most suited for that purpose. We report here the generation of syngeneic Malds generated against the murine monoclonal immunoglobulin T84.66 (Ab1), which defines a single epitope on the protein moiety of the carcinoembryonic antigen (CEA). This antigenic determinant is unique to CEA, as it is absent in other members of the CEA gene family that are expressed in a variety of normal human tissues, including granulocytes. The Mald 6G6.C4 (Ab2) exhibits the immunochemical features of an internal image antibody mimicking the epitope recognized by the idiotype T84.66. In enzyme immunoassays the binding of Ab1 to Ab2 is completely inhibited by CEA. In addition, Mald 6G6.C4 only binds to native but not to the denatured or reduced idiotype. Immunization of rabbits with F(ab')2-fragments of 6G6.C4 results in antisera (Ab3) that show specificity to CEA in both binding and inhibition enzyme immunoassays as well as in Western blots. Finally, Ab3 did not detect NCA, a major CEA-related glycoprotein in Western blots, either in a purified form or in a crude tumor extract, indicating a high specificity of the anti-anti-idiotypic response. In summary, these immunochemical data show that the monoclonal anti-idiotype 6G6.C4 can functionally mimic a CEA-specific epitope in rabbits and may do so in humans. Therefore, this antibody may have a clinical potential as a network antigen for active immune therapy in patients suffering from CEA-positive carcinomas.

Membrane association and shedding of the GPI-anchored Ca-MOv18 antigen in human ovary carcinoma cells

The antigen recognized by the MOv18 MAb (Ca-MOv18) was recently shown to be a glycosylphosphatidylinositol (GPI)-linked protein. In this report we show that GPI-anchorage is not limited to IGROVI cells nor to other ovary carcinoma cell lines, but Ca-MOv18 was also found to be sensitive to phosphatidylinositol-specific phospholipase C (PI-PLC) treatment on fresh ovarian cancer cells. Furthermore, we found a heterogeneous sensitivity of Ca-MOv18 to PI-PLC cleavage, not only among the different cells studied but also in different experiments performed on the same cell line, during extended periods of time in culture. Sensitivity to PI-PLC cleavage was determined by immunofluorescence on live cells and by double-determinant radioimmunoassay of the antigen released in the supernatant. The specificity of the PI-PLC cleavage was demonstrated as follows: (a) TX114 solubilized Ca-MOv18 shifts from the detergent to the aqueous phase after treatment with PI-PLC; (b) on membrane preparations, PI-PLC specifically released a fraction of the antigen, which is distinct from the weakly associated form released by high-salt treatment; (c) Ca-MOv18 from IGROVI expressed the cross-reacting determinant (CRD), which is characteristic of GPI-linked molecules. The absence of CRD expression on the spontaneously released protein and the possibility of artificially inducing antigen shedding during the resynthesis of Ca-MOv18 which follows bacterial PI-PLC treatment are interesting points which need to be further investigated in order to understand the physiology of the Ca-MOv18 tumor antigen.

A predicted three-dimensional structure for the carcinoembryonic antigen (CEA)

A three-dimensional model for the carcinoembryonic antigen (CEA) has been constructed by knowledge-based computer modelling. Each of the seven extracellular domains of CEA are expected to have immunoglobulin folds. The N-terminal domain of CEA was modelled using the first domain of the recently solved NMR structure of rat CD2, as well as the first domain of the X-ray crystal structure of human CD4 and an immunoglobulin variable domain REI as templates. The remaining domains were modelled from the first and second domains of CD4 and REI. Link conformations between the domains were taken from the elbow region of antibodies. A possible packing model between each of the seven domains is proposed. Each residue of the model is labelled as to its suitability for site-directed mutagenesis.

Assembly and analysis of cosmid contigs in the CEA-gene family region of human chromosome 19

The carcinoembryonic antigen (CEA)-like genes are members of a large gene family which is part of the immunoglobulin superfamily. The CEA family is divided into two major subgroups, the CEA-subgroup and the pregnancy-specific glycoprotein (PSG)-subgroup. In the course of an effort to develop a set of overlapping cosmids spanning human chromosome 19, we identified 245 cosmids in a human chromosome 19 cosmid library (6-7X redundant) by hybridization with an IgC-like domain fragment of the CEA gene. A fluorescence-based restriction enzyme digest fingerprinting strategy was used to assemble 212 probe-positive cosmids, along with 115 additional cosmids from a collection of approximately 8,000 randomly selected cosmids, into five contigs. Two of the contigs contain CEA-subgroup genes while the remaining three contigs contain PSG-subgroup genes. These five contigs range in size from 100 kb to over 300 kb and span an estimated 1 Mb. The CEA-like gene family was determined by fluorescence in situ hybridization to map in the q13.1-q13.2 region of human chromosome 19. Analysis of the two CEA-subgroup contigs provided verification of the contig assembly strategy and insight into the organization of 9 CEA-subgroup genes.

Long-range chromosomal mapping of the carcinoembryonic antigen (CEA) gene family cluster

A long-range physical map of the carcinoembryonic antigen (CEA) gene family cluster, which is located on the long arm of chromosome 19, has been constructed. This was achieved by hybridization analysis of large DNA fragments separated by pulse-field gel electrophoresis and of DNA from human/rodent somatic cell hybrids, as well as the assembly of ordered sets of cosmids for this gene region into contigs. The different approaches yielded very similar results and indicate that the entire gene family is contained within a region located at position 19q13.1-q13.2 between the CYP2A and the D19S15/D19S8 markers. The physical linkage of nine genes belonging to the CEA subgroup and their location with respect to the pregnancy-specific glycoprotein (PSG) subgroup genes have been determined, and the latter are located closer to the telomere. From large groups of ordered cosmid clones, the identity of all known CEA subgroup genes has been confirmed either by hybridization using gene-specific probes or by DNA sequencing. These studies have identified a new member of the CEA subgroup (CGM8), which probably represents a pseudogene due to the existence of two stop codons, one in the leader and one in the N-terminal domain exons. The gene order and orientation, which were determined by hybridization with probes from the 5' and 3' regions of the genes, are as follows: cen/3'-CGM7-5'/3'-CGM2-5'/5'-CEA-3'/5'-NCA-3'/5'-CGM1- 3'/3'-BGP-5'/3'- CGM9-5'/3'-CGM6-5'/5'-CGM8-3'/PSGcluster/qter.

The pregnancy-specific glycoprotein family of the immunoglobulin superfamily: identification of new members and estimation of family size

The members of the carcinoembryonic antigen (CEA)/pregnancy-specific glycoprotein (PSG) gene family have a characteristic N-terminal domain that is homologous to the immunoglobulin variable region. We have estimated the size of the PSG subfamily by identification of N-domain exons from isolated genomic clones and from total genomic DNA through PCR amplification and DNA sequence determination. The PSG subfamily contains at least 11 different genes. For 7 of these, two DNA sequences differing from each other in 1 to 4 nucleotides were detected. Most likely, they represent different alleles. They are PSG1, PSG2, PSG3, PSG4, PSG5, PSG6, PSG7, PSG8, PSG11, PSG12, and PSG13. Six of the N-domain sequences described here are new. All of the PSGs except PSG1, PSG4, and PSG8 contained the arginine-glycine-aspartic acid sequence at position 93-95 corresponding to the complementarity determining region 3 of immunoglobulin. Parsimony analysis of 24 CEA and PSG sequences using 12 members of the immunoglobulin gene superfamily as outgroups to root the family tree shows that the N-domain of the CEA group genes evolved in one major branch and the PSG group genes in the other.

Characterization of a major member of the rat pregnancy-specific glycoprotein family

The gene encoding the human pregnancy-specific glycoprotein (PSG) belongs to a gene subfamily, comprised of the carcinoembryonic antigen (CEA) and PSG subgroups, within the immunoglobulin superfamily. To study the functional roles of PSG during development in an animal model, we isolated and characterized a near full-length cDNA (rnCGM6) encoding a PSG-related protein from a rat placental cDNA library. rnCGM6 is 2,068 bp in length and contains an open reading frame that encodes a 475-amino-acid polypeptide with a predicted molecular mass of 53 kD. The 5' noncoding sequence is 173 nucleotides, and primer-extension experiments demonstrate that the transcriptional initiation site is located 22-24 nucleotides further upstream. The 3' noncoding sequence contains 470 nucleotides which is followed by a poly(A) tail. In contrast to human PSGs, which contain one immunoglobulin variable-like and two to three immunoglobulin constant-like protein domains, rnCGM6 contains three immunoglobulin variable-like domains and one immunoglobulin constant-like domain. rnCGM6 contains six potential N-linked glycosylation sites and, in its carboxyl-terminal domain, a tyrosine protein kinase phosphorylation site. The tyrosine phosphorylation site is conserved among all rat and human PSG members. rnCGM6 hybridized with a major 2.5-kb and two minor 3.0- and 3.5-kb mRNAs, all primarily expressed in the rat placenta. Ribonuclease protection analysis, using probes specific to the 5', middle, and 3' regions of rnCGM6, and the 5' region of a previously identified cDNA, rnCGM1, mainly yielded fully-protected fragments indicating relatively low sequence similarity among rat PSG-related proteins. Northern hybridization and ribonuclease protection assays also suggest that rnCGM6 may be the major PSG member in rat.

Tyrosine phosphorylation of biliary glycoprotein, a cell adhesion molecule related to carcinoembryonic antigen

Biliary-glycoprotein (BGP), a cell adhesion molecule related to carcinoembryonic antigen (CEA), has been shown to exist as several alternatively spliced isoforms. Here we show that BGPa and BGPb are phosphorylated in the chronic myelogenous leukaemia cell line KG-1, which constitutively expresses several BGP isoforms, and Chinese hamster LR-73 cells transfected with the cDNAs encoding BGPa and BGPb. The phosphorylation can be augmented with the protein tyrosine phosphatase inhibitor ammonium vanadate and with TPA (an activator of protein kinase C). Phospho-amino acid analysis of phosphorylated BGPs demonstrated that phosphorylation occurs on serine, threonine and tyrosine residues. Phosphorylation reactions carried out in in vitro membrane preparations from KG-1 cells revealed a close association of BGP proteins with membrane associated protein tyrosine kinases. These observations suggest an association of BGP proteins with signal transduction molecules which is regulated by alternative splicing of the cytoplasmic domain.

Molecular characterization of the Schwann cell myelin protein, SMP: structural similarities within the immunoglobulin superfamily

The Schwann cell myelin protein (SMP), previously defined in quail and chick by a monoclonal antibody, is in vivo exclusively expressed by myelinating and nonmyelinating Schwann cells and oligodendrocytes. The isolation of the complete nucleotide sequence of SMP is reported here. The predicted polypeptide chain reveals that SMP is a transmembrane molecule of the immunoglobulin superfamily showing sequence similarities with several surface glycoproteins expressed in the nervous and immune systems. In spite of a 43.5% overall sequence identity between rat myelin-associated glycoprotein (MAG) and quail SMP, SMP does not seem to be the avian homolog of MAG, since their expression, regulation, and functions are significantly different. Unusual sequence arrangements shared by SMP, MAG, and two lymphoid antigens suggest the existence of a particular subgroup in the immunoglobulin superfamily.

Cloning of the mouse hepatitis virus (MHV) receptor: expression in human and hamster cell lines confers susceptibility to MHV

The cellular receptor for murine coronavirus mouse hepatitis virus (MHV)-A59 is a member of the carcinoembryonic antigen (CEA) family of glycoproteins in the immunoglobulin superfamily. We isolated a cDNA clone (MHVR1) encoding the MHV receptor. The sequence of this clone predicts a 424-amino-acid glycoprotein with four immunoglobulinlike domains, a transmembrane domain, and a short intracytoplasmic tail, MHVR1 is closely related to the murine CEA-related clone mmCGM1 (Mus musculus carcinoembryonic antigen gene family member). Western blot (immunoblot) analysis performed with antireceptor antibodies detected a glycoprotein of 120 kDa in BHK cells stably transfected with MHVR1. This corresponds to the size of the MHV receptor expressed in mouse intestine and liver. Human and hamster fibroblasts transfected with MHVR1 became susceptible to infection with MHV-A59. Like MHV-susceptible mouse fibroblasts, the MHVR1-transfected human and hamster cells were protected from MHV infection by pretreatment with monoclonal antireceptor antibody CC1. Thus, the 110- to 120-kDa CEA-related glycoprotein encoded by MHVR1 is a functional receptor for murine coronavirus MHV-A59.

Specific adhesion of carcinoembryonic antigen-bearing colorectal cancer cells to immobilized carcinoembryonic antigen

Recent characterization of the genomic structure of carcinoembryonic antigen (CEA) is consistent with that of a cellular adhesion molecule. To examine this function in colorectal cancer, the adherence of cell lines to microtiter wells coated with CEA and well-described adhesive molecules was determined. The CEA-positive cell line LoVo and the CEA-devoid cell line H-Meso-1 did not differ in adherence to the extracellular matrix proteins laminin, collagen and fibronectin, whereas LoVo cells adhered to CEA (10 micrograms/well) in a specific manner (43% bound cells vs. 1.5% bound cells with BSA or alpha-acidglycoprotein controls, P less than 0.01) while H-MESO-1 showed no adhesion to CEA (less than 0.6% bound cells). This adhesion of LoVo cells to CEA was not affected by co-incubation of cells with EDTA, sodium azide, or at 23 degrees C. However, the CEA to CEA adhesive interaction was inhibited by a monoclonal antibody directed against an epitope in the N-terminal domain of the CEA molecule, and decreased by enzymatic removal of CEA from the LoVo cell membrane. The extent of adhesion to immobilized CEA by four CEA-producing cell lines (LoVo, HT29, LS174T and LS174-S), correlated with membrane CEA expression as determined by FACS analysis. The results of these experiments add support to the concept that CEA may function as a specific homotypic cellular adhesion molecule for colorectal cancer cells.

The mouse biliary glycoprotein gene (Bgp): partial nucleotide sequence, expression, and chromosomal assignment

Transcripts related to the human carcinoembryonic antigen were found in mRNA isolated from both dimethylbenzanthracene-induced and mouse mammary tumor virus-induced mammary tumors. A cDNA library was prepared from a dimethylbenzanthracene-induced tumor, and a clone was isolated by hybridization with a human carcinoembryonic antigen cDNA probe. Its sequence, when compared to those of members of the human carcinoembryonic antigen gene family, was most homologous to the sequence of the human biliary glycoprotein (BGP) gene. Thus, this clone appears to encode a portion of the mouse biliary glycoprotein gene. Southern blot analysis of EcoRI-digested mouse cellular DNA with this probe detected four restriction fragments, all of which appear to be located on mouse chromosome 7, Northern blot analysis using the mouse probe demonstrated that related mRNA species were expressed in some normal adult mouse tissues.

Binding of nonspecific cross-reacting antigen, a granulocyte membrane glycoprotein, to Escherichia coli expressing type 1 fimbriae

Nonspecific cross-reacting antigen (NCA) is a well-characterized membrane glycoprotein on granulocytes, macrophages, and lung epithelium. Structural studies at the protein and genomic levels have revealed that NCA is a member of the immunoglobulin supergene family, and hybridization studies showed that the transcript level of NCA is induced by treatment with gamma interferon. These studies, as well as the expression of NCA on granulocytes, suggest a role for NCA in immune response. For a first step in studying this possible role, we have examined the binding of two glycoforms of NCA designated NCA-50 (Mr, 50,000) and TEX-75 (Mr, 75,000). Here we report the results from binding assays which demonstrate carbohydrate-mediated binding of Escherichia coli expressing type 1 fimbriae and of isolated type 1 fimbriae to NCA-50. TEX-75 did not bind to the purified fimbriae but bound slightly to the bacterial strain. Inhibition studies showed that the binding to NCA-50 involved interaction of mannose moieties on NCA-50 and lectins on the fimbriae. The binding of NCA-50 to bacterial fimbriae was confirmed by electron microscopy studies, using immunolabeling techniques. In addition, we show that the surface expression of NCA-50 (and presumably of other NCA species) on isolated polymorphonuclear leukocytes is increased following activation with the bacterial peptide formylmethionyl-leucyl-phenylalanine, consistent with a role for NCA in immune response.

Receptor for mouse hepatitis virus is a member of the carcinoembryonic antigen family of glycoproteins

The receptor for mouse hepatitis virus (MHV), a murine coronavirus, is a 110- to 120-kDa glycoprotein on intestinal brush border membranes and hepatocyte membranes. The N-terminal 25-amino acid sequence of immunoaffinity-purified MHV receptor was identical to the predicted mature N termini of two mouse genes related to human carcinoembryonic antigen (CEA) and was strongly homologous to the N termini of members of the CEA family in humans and rats. Polyclonal antibodies to human CEA recognized the immunoaffinity-purified MHV receptor and the MHV receptor in liver membranes and intestinal brush border membranes from MHV-susceptible mouse strains. In membranes from MHV-resistant SJL/J mice, the anti-CEA antibodies recognized a homologous glycoprotein that failed to bind MHV. The MHV receptor glycoprotein was detected in membranes of BALB/c colon, small intestine, and liver, which are the principal targets for MHV replication in vivo. The MHV receptor glycoprotein resembled members of the human CEA family in molecular weight, acidic pI, extensive glycosylation, solubility in perchloric acid, and tissue distribution. Thus, the MHV receptor is, to our knowledge, the first member of the CEA family of glycoproteins to be identified as a virus receptor.

Three novel molecular forms of biliary glycoprotein deduced from cDNA clones from a human leukocyte library

Three cDNA clones that encode three novel variants of biliary glycoprotein a (BGPa), a glycoprotein belonging to the CEA gene family, were identified in a human leukocyte cDNA library. The domain structures of the predicted proteins of the three clones W211, W233 and W239 are N-A1-B1-A2, N-A1-B1 and N-A1-B1-C, respectively; they lack the transmembrane and cytoplasmic domains that exist in the four BGP species (BGPa, BGPb, BGPc and BGPd) previously reported. Their sequences from N to B1 or to A2 are virtually identical to those of BGPa-d. Comparison with the genomic sequence for BGPa-d suggested that these three new BGP variants as well as BGPa-d are generated from the same single gene by alternative splicing of RNA.

Identification of membrane antigens in granulocytes and colonic carcinoma cells by a monoclonal antibody specific for biliary glycoprotein, a member of the carcinoembryonic antigen family

The gene coding for 'biliary glycoprotein (BGP)' is a member of the carcinoembryonic antigen (CEA) gene family. A monoclonal antibody (MAb) was induced against a BGP-preparation isolated from human bile. The antibody did not crossreact with the carcinoembryonic antigen (CEA) and different non-specific crossreacting antigens. The anti-BGP MAb was used to identify BGP-related antigens in membrane extracts from granulocytes and the colonic carcinoma cell line HT-29. In granulocyte membranes, a single antigen of Mr 160,000 was bound. In membranes from HT-29 cells, a main antigen of Mr 85,000 was present. At high antigen concentration, an additional antigen of Mr 115,000 was identified. Since several transcripts of the BGP gene have been identified, the different BGP related antigens are probably products of alternatively spliced mRNAs.

Carcinoembryonic antigen gene family: molecular biology and clinical perspectives

The carcinoembryonic antigen (CEA) gene family belongs to the immunoglobulin super-gene family and can be divided into two main subgroups based on sequence comparisons. In humans it is clustered on the long arm of chromosome 19 and consists of approximately 20 genes. The CEA subgroup genes code for CEA and its classical crossreacting antigens, which are mainly membrane-bound, whereas the other subgroup genes encode the pregnancy-specific glycoproteins (PSG), which are secreted. Splice variants of individual genes and differential post-translational modifications of the resulting proteins, e.g., by glycosylation, indicate a high complexity in the number of putative CEA-related molecules. So far, only a limited number of CEA-related antigens in humans have been unequivocally assigned to a specific gene. Rodent CEA-related genes reveal a high sequence divergence and, in part, a completely different domain organization than the human CEA gene family, making it difficult to determine individual gene counterparts. However, rodent CEA-related genes can be assigned to human subgroups based on similarity of expression patterns, which is characteristic for the subgroups. Various functions have been determined for members of the CEA subgroup in vitro, including cell adhesion, bacterial binding, an accessory role for collagen binding or ecto-ATPases activity. Based on all that is known so far on its biology, the clinical outlook for the CEA family has been reassessed.

The effect of interferon treatment on 14 human colorectal cancer cell lines: growth and carcinoembryonic antigen secretion in vitro

The effect of interferons (IFNs) on growth inhibition and carcinoembryonic antigen (CEA) secretion by 14 established human colorectal carcinoma cell lines was studied in vitro. The cell lines were characterized by morphologic differentiation, level of CEA production, and rate of growth. All cell lines were treated in vitro with recombinant human IFN (alpha, beta, and gamma) and the effect of treatment on growth rate and CEA secretion determined. Each cell line exhibited an individual pattern of growth inhibition that was independent of degree of differentiation, level of CEA production, and rate of growth. IFN-beta treatment did not increase CEA secretion in any of the cell lines studied. IFN-alpha and IFN-gamma resulted in increased CEA production (2- to 81-fold increase) primarily in the moderately to well-differentiated cell lines. IFN-gamma was a more potent inducer of enhanced CEA secretion than IFN-alpha. The more poorly differentiated cell lines did not produce CEA and could not be induced to do so by any of the IFNs.

Induction of carcinoembryonic-antigen-gene expression in human colorectal carcinoma by sodium butyrate

The effect of sodium butyrate on the expression of the carcinoembryonic-antigen (CEA) gene was studied in two poorly differentiated colorectal-carcinoma cell lines (Clone-A and MIP-101) and in one well-differentiated cell line (LS-174T); A.T.C.C. no. CCL 188). Northern-blot and dot-blot analyses indicated a steady increase in CEA mRNA from day 4 to a maximal level by day 14 after these cells were exposed to 2 mM-sodium butyrate. Studies using nuclear run-off assays followed by dot-blot hybridization to a partial CEA cDNA clone demonstrated that specific increases in gene transcription rates (3-fold in MIP-101, 4-fold in LS-174T and 6-fold in Clone-A) are not sufficient to account for the observed increases in CEA mRNA abundance. Further studies showed that CEA-specific transcripts have a half-life of about 60-80 min, and treatment with sodium butyrate increased the stability of CEA-specific transcripts to about 340 min in LS-174T cells and to about 500 min in Clone-A cells. We conclude that the induction of the CEA-gene expression by sodium butyrate in colorectal-cancer cells is mediated by both transcriptional and post-transcriptional mechanisms, with CEA mRNA stability as one of the major check-points.

Cloning of the complete gene for carcinoembryonic antigen: analysis of its promoter indicates a region conveying cell type-specific expression

Carcinoembryonic antigen (CEA) is a widely used tumor marker, especially in the surveillance of colonic cancer patients. Although CEA is also present in some normal tissues, it is apparently expressed at higher levels in tumorous tissues than in corresponding normal tissues. As a first step toward analyzing the regulation of expression of CEA at the transcriptional level, we have isolated and characterized a cosmid clone (cosCEA1), which contains the entire coding region of the CEA gene. A close correlation exists between the exon and deduced immunoglobulin-like domain borders. We have determined a cluster of transcriptional starts for CEA and the closely related nonspecific cross-reacting antigen (NCA) gene and have sequenced their putative promoters. Regions of sequence homology are found as far as approximately 500 nucleotides upstream from the translational starts of these genes, but farther upstream they diverge completely. In both cases we were unable to find classic TATA or CAAT boxes at their expected positions. To characterize the CEA and NCA promoters, we carried out transient transfection assays with promoter-indicator gene constructs in the CEA-producing adenocarcinoma cell line SW403, as well as in nonproducing HeLa cells. A CEA gene promoter construct, containing approximately 400 nucleotides upstream from the translational start, showed nine times higher activity in the SW403 than in the HeLa cell line. This indicates that cis-acting sequences which convey cell type-specific expression of the CEA gene are contained within this region.

Cloning of the complete gene for carcinoembryonic antigen: analysis of its promoter indicates a region conveying cell type-specific expression

Carcinoembryonic antigen (CEA) is a widely used tumor marker, especially in the surveillance of colonic cancer patients. Although CEA is also present in some normal tissues, it is apparently expressed at higher levels in tumorous tissues than in corresponding normal tissues. As a first step toward analyzing the regulation of expression of CEA at the transcriptional level, we have isolated and characterized a cosmid clone (cosCEA1), which contains the entire coding region of the CEA gene. A close correlation exists between the exon and deduced immunoglobulin-like domain borders. We have determined a cluster of transcriptional starts for CEA and the closely related nonspecific cross-reacting antigen (NCA) gene and have sequenced their putative promoters. Regions of sequence homology are found as far as approximately 500 nucleotides upstream from the translational starts of these genes, but farther upstream they diverge completely. In both cases we were unable to find classic TATA or CAAT boxes at their expected positions. To characterize the CEA and NCA promoters, we carried out transient transfection assays with promoter-indicator gene constructs in the CEA-producing adenocarcinoma cell line SW403, as well as in nonproducing HeLa cells. A CEA gene promoter construct, containing approximately 400 nucleotides upstream from the translational start, showed nine times higher activity in the SW403 than in the HeLa cell line. This indicates that cis-acting sequences which convey cell type-specific expression of the CEA gene are contained within this region.

Carcinoembryonic antigen gene family members in submandibular salivary gland: demonstration of pregnancy-specific glycoproteins by cDNA cloning

We have recently shown that human submandibular salivary gland and saliva contain a number of glycoproteins belonging to the carcinoembryonic antigen (CEA) gene family. The members of the CEA family can be divided into the CEA subgroup and the pregnancy specific beta 1 glycoprotein (PSG) subgroup. The latter glycoproteins are abundant in placenta and fetal liver. Here we report that PSG's are expressed in normal adult submandibular salivary gland. Thus, cDNA cloning and sequencing gave two clones (SG5 and SG9) which coded for glycoproteins with a domain arrangement of N-A1-A2-B2-C and a third clone (SG8) which coded for a glycoprotein with a domain arrangement of N-A1-B2-C. SG5 is identical to PSG3, and SG9 to PSG1d, while SG8 most probably corresponds to PSG2. The 3' untranslated regions of the different members of the PSG subgroup contain highly homologous segments, suggesting a common evolutionary origin.

Transcription of biliary glycoprotein I gene in malignant and non-malignant human liver tissues

Biliary glycoprotein I (BGP I) is a member of carcinoembryonic antigen (CEA) gene family consisting of at least 11 related genes. The transcription of BGP I gene was analysed in malignant and non-malignant human liver tissues with a 396-bp 3'-untranslated region probe from a cDNA clone 4-13 which was newly isolated from an adult human colon cDNA library. Among 21 tissue samples from 14 patients with hepatocellular carcinoma, 16 samples were clearly shown to express a single 3.9-kb message. This message was also found in the hepatoma cell line HuH-7. When the malignant tissues were compared to the non-malignant ones for the intensity of the band, no significant difference was observed. mRNAs of CEA and non-specific cross-reacting antigen (NCA) were not detected in 5 samples which were shown to have the message of the BGP I gene. These data suggest that the human hepatocyte and its malignant transformant produce BGP I, and that this could correspond to the cross-reacting antigen previously detected in the liver.

Identification of a new carcinoembryonic antigen (CEA) family member in human fetal liver--cloning and sequence determination of pregnancy-specific glycoprotein 7

The carcinoembryonic antigen gene family consists of the CEA- and the Pregnancy-Specific Glycoprotein- (PSG) subfamilies. Human fetal liver express several PSGs. Here we report cloning and sequencing of a new PSG subfamily member, PSG7. It is the fifth type of PSG found in fetal liver. PSG7 has the N-A1-A2-B2-C domain arrangement. Unlike other PSGs the N-terminal of PSG7 is unblocked. PSG7 has a cysteine in the C-terminal domain, which may allow dimerization. Variability analysis according to Wu and Kabat reveals that the region in the N-domain corresponding to complementarity determining region 3 of immunoglobulin is different between PSG subfamily members. Many members, including PSG7, contain the RGD sequence in this region. The CD2 region as well as two other short sequences (in N and A1 domains respectively) also show some variability. The function of PSGs is probably linked to the N-domain and the CDR2- and CD3-like regions are most likely responsible for ligand binding.

Structure, evolution and chromosomal localization of the human pregnancy-specific beta 1 glycoprotein gene family

Cloning and sequencing of the cDNA of a new member of the PS beta G gene family is reported. Comparison of the sequence with those of other PS beta G cDNAs reveals a remarkable conservation of their sequence (greater than 90%) and of their general structural organization: an NH2 domain is followed by 93- and 85-residue Ig-like domains termed A and B, respectively. Most PS beta G contain two domains A in tandem followed by a single domain B. In some PS beta G members described here, alternative splicing skips the AI domain in some transcripts, yielding two- and three-domain variants, respectively. Individual PS beta G members have specific short carboxyl domains displaying little sequence conservation. The PS beta G family is closely related to the CEA gene family. A detailed comparison of the sequence of both families is given and used to construct an evolutionary tree, using the method of Li, Wu, and Luo (1985, Mol. Biol. Evol. 2: 150-174). Computation of the number of substitutions of synonymous (Ks) and nonsynonymous (Ka) sites and of the Ks/Ka ratio suggests that the PS beta G gene family appeared concomitantly with the expansion of the placental mammals and belongs to the class of rapidly evolving genes. Very little selective pressure has been exerted on the body of the molecules, especially on domain A. The analysis also suggests that PS beta G genes encoding different carboxyl domains would have been positively selected and fixed during the evolution. The PS beta G gene family was assigned to chromosome 19, which also carries the CEA genes.

Characterization of cDNAs of the human pregnancy-specific beta 1-glycoprotein family, a new subfamily of the immunoglobulin gene superfamily

Three highly homologous cDNAs encoding human pregnancy-specific beta 1-glycoprotein (SP1) were isolated from a human placental cDNA library. These cDNAs share greater than 90% nucleotide homology in their coding sequences, and greater than 79% of the encoded amino acids are homologous. Proteins encoded by these cDNAs are very similar to members of the carcinoembryonic antigen family and contain repeating domains, conserved disulfide bridges, and beta-sheet structure typical of the immunoglobulin gene superfamily. However, the high degree of sequence homology and relatively lesser degree of glycosylation among the SP1 proteins suggest that they exist as a unique family instead of being members of the CEA family. Both soluble and potentially membrane-bound forms of SP1 proteins were present in the placenta. Northern blot analysis using specific probes confirmed the expression of multiple mRNA species in human term placenta.

The human pregnancy-specific glycoprotein genes are tightly linked on the long arm of chromosome 19 and are coordinately expressed

The pregnancy-specific glycoprotein (PSG) genes encode a group of proteins which are found in large amounts in placenta and maternal serum. In situ hybridization analyses of metaphase chromosomes reveal that all the human pregnancy-specific glycoprotein (PSG) genes are located on the long arm of chromosome 19 (19q13.2-13.3), overlapping the region containing the closely-related carcinoembryonic antigen (CEA) gene subgroup. Higher resolution analyses indicate that the PSG genes are closely linked within an 800kb SacII restriction endonuclease fragment. This has been confirmed through restriction endonuclease mapping and DNA sequence analyses of isolated genomic clones, which show that at least some of these genes are located in very close proximity. Further, these studies have helped to identify a new member of the PSG gene subfamily (PSG7). DNA/RNA hybridization analyses, using gene-specific oligonucleotide probes based on published sequences, showed that five from six PSG genes tested are coordinately transcribed in the placenta. Due to the close proximity of these genes and their coordinated expression pattern, common transcriptional regulatory elements may exist.

Characterization of a cDNA clone encoding a new species of the nonspecific cross-reacting antigen (NCA), a member of the CEA gene family

To clarify the molecular structures of the nonspecific cross-reacting antigens (NCAs) produced by human granulocytes, we cloned cDNAs from libraries of normal white blood cells. A clone, NCA-W272, was found to code a protein similar to NCA of tumor cells. The protein consisted of a signal peptide (34 aa), domain-N (108 aa), -A1 (92 aa), -B1 (86 aa) and -M (29 aa). Similarity of the amino acid sequence of each domain to that of the tumor NCA was 72, 92, 76 and 79%, respectively. COS-1 cells transfected with an expression vector carrying the cDNA synthesized a 70 kDa glycoprotein, which was reactive with anti-NCA antibody and released from cell surface by phosphatidylinositol-specific phospholipase C. Thus the clone NCA-W272 was indicated to encode a new species of NCA distinct from the tumor NCA.

Carcinoembryonic antigen functions as an accessory adhesion molecule mediating colon epithelial cell-collagen interactions

We have previously shown that a human colon carcinoma cell line (SW1222) expresses a collagen receptor recognizing the Arg-Gly-Asp tripeptide sequence found in collagen. This receptor mediates the cellular attachment to collagen and, subsequently, the glandular differentiation seen in a three-dimensional collagen gel culture. In a search to identify cell surface molecules mediating the adhesion and differentiation of SW1222 cells, we have screened a panel of monoclonal antibodies recognizing epithelial cell surface determinants for their ability to inhibit the collagen binding of SW1222 cells. We have found that four monoclonal antibodies recognizing the 180-kDa carcinoembryonic antigen (CEA) glycoprotein and other members of the CEA family inhibited (up to 87%) the binding of SW1222 cells to type I collagen matrix. Using a cell attachment assay, we have not detected any direct collagen binding of either purified CEA or another CEA-expressing human colon carcinoma cell line (LS174T). These data suggest that CEA is not a collagen-binding protein itself but is likely to be associated with the functional Arg-Gly-Asp collagen receptor expressed by SW1222 cells. We suggest that CEA may function as an accessory molecule, controlling the functional activity of the SW1222 collagen receptor.

Carcinoembryonic antigen: function in metastasis by human colorectal carcinoma

Carcinoembryonic antigen (CEA) is a glycoprotein that has been useful as a tumor marker to predict recurrence in gastrointestinal malignancies, but whose biological function has not been elucidated. With the recent evidence that CEA is a member of the immunoglobulin supergene family, CEA may be involved in intercellular recognition and binding. This review examines the role that CEA plays in the development of metastases by colorectal carcinoma.

cDNA cloning demonstrates the expression of pregnancy-specific glycoprotein genes, a subgroup of the carcinoembryonic antigen gene family, in fetal liver

The pregnancy-specific glycoprotein (PSG) genes constitute a subgroup of the carcinoembryonic antigen (CEA) gene family. Here we report the cloning of four cDNAs coding for different members of the PSG family from a human fetal liver cDNA library. They are derived from three closely related genes (PSG1, PSG4 and PSG6). Two of the cDNA clones represent splice variants of PSG1 (PSG1a, PSG1d) differing in their C-terminal domain and 3'-untranslated regions. All encoded proteins show the same domain arrangement (N-RA1-RA2-RB2-C). Transcripts of the genes PSG1 and PSG4 could be detected in placenta by hybridization with gene-specific oligonucleotides. Expression of cDNA in a mouse and monkey cell line shows that the glycosylated PSG1a protein has a Mr of 65-66 kD and is released from the transfected cells. Sequence comparisons in the C-terminal domain and the 3'-untranslated regions of CEA/PSG-like genes suggests a complex splicing pattern to exist for various gene family members and a common evolutionary origin of these regions.

Limited improvement of tumour diagnosis by the simultaneous determination of carcinoembryonic antigen (CEA) and of a tumour-associated CEA-related antigen of Mr 128,000 in serum

Two biotin-avidin based enzyme immunoassays were developed using three monoclonal anti-CEA antibodies with distinct epitope- and antigen-specificities. A broadly cross-reactive monoclonal anti-CEA antibody (T84.1) was immobilized on a solid support. Either monoclonal antibody T84.66 or CEA.11 was used as the second, biotin-labeled monoclonal antibody. Both antibodies do not cross-react with normal granulocytes or bile canaliculi. Monoclonal antibody CEA.11 binds to CEA and to an antigen with a relative molecular mass of 128,000 present in extracts from solid carcinomas. Monoclonal antibody T84.66 does not cross-react with the antigen of Mr 128,000. After adsorption of tumour extracts to a CEA-specific immunosorbent, residual activity was measurable by the CEA.11 assay, but not by the T84.66 assay. Serum samples (n = 726) from patients with malignant and non-malignant disease, as well as from healthy volunteers, were analysed by both immunoassays and by two commercial CEA immunoassays. In comparison with the T84.66 assay and the commercial assays, the CEA.11 assay did not significantly increase the sensitivity or specificity of tumour diagnosis.

Human pregnancy-specific beta 1-glycoproteins are coded within chromosome 19

We have isolated and characterized cDNAs that code for apoproteins having amino acid sequences highly similar to pregnancy-specific beta 1-glycoproteins (PS beta G). cDNAs coding for PS beta Gs, as well as the cDNA clone reported here, are members of the carcinoembryonic antigen (CEA) gene family. The previous localization of CEA-related genes to human chromosome 19, and the high level of DNA sequence conservation in the CEA family, suggested that the PS beta G genes are also located on this chromosome. We demonstrate here that chromosome 19 is indeed the site of PS beta G sequences. Our finding is in contrast to the recently reported indication that pregnancy-specific glycoproteins are encoded in chromosomes X and 6.